Occasionally, a number of technical advances come together to give a quantum leap forward. This has occurred recently as a result of three factors - the increased density of components on an integrated circuit; the successful fabrication of fault- tolerant complete integrated circuit wafers; and a new patented approach to structuring these wafers by Catt called the Kernel Logic invention. The result is that the latent, explosive power of semiconductor technology can be unleashed - one million computers working together in an array to solve large, complex problems at high speed.
The first Wafer Scale Integration product, a solid state disc called WaferStack, came to market in 1989, based on "Catt Spiral". We can now advance to a WSI array processor, The Kernel Machine, with one million processors giving one million million operations per second (= one teraflop).
The Kernel Machine, built from an array of 100 wafers, retails for £500,000. The external controller draws a map of good and bad chips, and devises a strategy for building a volatile, perfect square two dimensional array of 1,000,000 processing elements (PE's) out of a larger, imperfect array.
The Kernel Project Plan.
The project costs £20 millions spread over four years. We build and deliver four machines, including all general purpose software, plus turnkey software for one application. Revenue begins to come in in the fifth year. If an initial £40M outlay plus interest (25% compound) were to be clawed back during the next four years, then we would need to claw back clear profit of £20M p.a.
A proper figure for profit in this market would be high; 20% of retail price. So, assuming a cost overrun of 100%, a sufficient turnover during those four years is £100M (or $200M) p.a.
The Kernel Machine is the standard array processor of the future, as the von Neumann single processor machine was the standard during the twentieth century. In Electronics and Wireless World, March 1989, I have discussed the many applications for the Kernel Machine with 1000 x 1000 PE's. These applications most fashionably include the only way to do world-wide modelling for global warming. 75% of Cray applications - Automotive, Aerospace, Petroleum, Chemistry, Meteorology and Earth Systems - are appropriate. Cray's turnover of $0.8Bn out of a world supercomputer market of £2Bn represents a default market, only indicating the much larger potential $12Bn market when the correct machine (Kernel) arrives.
Extracted from Catt's paper for the IEE Colloquium on 28 May 1991. Additions from "The Kernel Logic Machine", Electronics and Wireless World, March 1989, page 254. Patent nos. Europe 88905982.0; USA 5,055,774 Oct. 8, 1991. Also Japan.
Ivor Catt, Kernel Consultants,
121 Westfields, St. Albans AL3 4JR, England. 0727 864257. sept93.
Top Of Page